基于计算机技术的中医面色诊信息采集与识别研究  

Computer Based Information Acquisition and Recognition for Complexion Diagnosis

作　　者：李福凤[1] 邸丹[1] 王忆勤[1] 钱鹏[1] 郑晓燕[1] 刘国萍[1] 李晓强[2] 

机构地区：[1]上海中医药大学中医证实验室,上海201203 [2]上海大学计算机工程与科学学院,上海200072

出　　处：《世界科学技术-中医药现代化》2008年第6期71-76,共6页Modernization of Traditional Chinese Medicine and Materia Medica-World Science and Technology

基　　金：国家自然基金项目(30600796):基于多种信息处理技术的面色诊信息自动识别研究,负责人:李福凤;上海市第三期重点学科(S30302):中医诊断学,负责人:王忆勤

摘　　要：目的:探索中医面色诊的规范化采集环境和客观识别方法,实现中医面色诊的计算机自动识别。方法:在中医理论指导下,比较各种人造光源,建立面色诊采集环境和临床面色判读量表;采集典型面色样本,结合图像处理、模式识别、人工智能等多种信息处理方法对面部颜色信息进行识别。结果:①研制了符合中医特点的以LED光源为照明光源的面色诊采集暗箱;②确定了符合中医理论的,临床专家认可的中医面部五色分类判读量表;③建立了面色诊颜色客观自动识别的方法,确定各类面色在色彩空间中的范围,初步研制中医面色识别分析系统。结论:该方法实现了中医面色诊的计算机自动识别,克服了传统中医望面色无客观量化标准和存在主观性的不足,增强了中医临床信息采集的规范化和准确性,提高了辨证诊断的科学性。The paper discusses a possible standardized environment where appropriate methods can be used for complexion acquisition and objective recognition, with the help of a computer based automatic complexion differentiation and diagnosis system. Researchers have established an observational scale for complexion acquisition and clinical judgment, based on TCM theory. They compared the data with some artificial light sources, and collected the typical complexion samples for recognition, by combining different message processing methods, including imagery processing, pattern recognition, and artificial intelligence. The study has come up with the following results： （1） LED light source for the first time to replace the sunlight. An optimal setup of complexion collecting environment was realized; is used （2） five color complexion classification observational scales were defined, in line with TCM theories and clinical practice; （3） an automatic complexion discernment and analysis system was established, and the full complexion color range was defined. A TCM complexion recognition analysis system has been basically established. It is believed that the method has made the computer aided automatic complexion recognition possible. The method has overcome the weakness of the traditional complexion recognition method, featured with no objective quantification standards and certain subjectivity. It enhances TCM clinical information acquisition standardization and the accuracy of dialectical diagnosis.

关 键 词：面色诊 采集环境 图像处理 计算机技术 

分 类 号：R445[医药卫生—影像医学与核医学] R22[医药卫生—诊断学]